There was some talk in the context of the new protocol to do some
enhancements in the area of the "binary" format to make it useful across
different architectures etc.
The SQL standard defines a concept called "transform groups", which are
basically pairs of input/ouput functions that can be switched between.
The standard talks about different transform groups for different host
languages, so this essentially selects between different binary output
formats.
I think this would fit naturally with many things we are doing and want to
do. We could have a (default) "text" transform group, a "binary little
endian" transform group, a "binary big endian" transform group".
Transform groups could map more sophisticated data types into native
programming language structures (think timestamp or inet). Users could
also plug in their own transform groups for existing data types if they
hate the input or output format. It would also reduce the pressure to add
more ad-hoc format-twiddling parameters like extra_float_digits or even
datestyle.
Is this something worth investigating?
--
Peter Eisentraut peter_e@gmx.net